Recurrence of autoimmune liver disease after liver transplantation: A systematic review



Recurrence of autoimmune liver disease in allografts has long been a topic of debate. We conducted a systematic review of the literature to examine the reported incidence of recurrence after liver transplantation of primary biliary cirrhosis (PBC), primary sclerosing cholangitis (PSC), and autoimmune hepatitis (AIH). The MEDLINE, EMBASE, and Cochrane electronic databases were used to identify articles. The inclusion criteria used were articles on patients with at least 90 days of posttransplantation follow-up, histologic criteria for diagnosis of PBC and AIH recurrence, radiologic or histologic criteria or both for diagnosis of PSC recurrence, and exclusion of other causes of liver disease causing similar histologic findings. Incidence in individual studies was combined to calculate the overall recurrence. Risk factors were analyzed whenever crude data were available. Funnel plots were used to assess publication bias. Out of 90 articles identified, 43 met criteria for systematic review (PBC, 16; PSC, 14; AIH, 13). The calculated weighted recurrence rate was 18% for PBC, 11% for PSC, and 22% for AIH. No difference was found in PBC and AIH recurrence by type of primary immunosuppression. There were not enough data to assess this issue in PSC studies. There was evidence of publication bias among PSC and AIH studies but not among PBC studies. In conclusion, recurrence of autoimmune liver disease after liver transplantation appears to be a real concern. As these patients are followed long-term, recurrence of disease may become the primary cause of morbidity. Liver Transpl 12:1813-1824, 2006. © 2006 AASLD.

Liver transplantation is a well-accepted treatment modality for autoimmune liver disease. Recurrence of primary biliary cirrhosis (PBC), primary sclerosing cholangitis (PSC), and autoimmune hepatitis (AIH) after liver transplantation has long been a subject of debate. Different authors have reported various experiences with the recurrence of these diseases in allografts.

Recurrence of PBC was first reported in 1982 by Neuberger et al.1 Recurrent disease was diagnosed on the basis of clinical features, abnormal liver test results, elevated antimitochondrial antibodies, and biopsy findings compatible with features of PBC. Because of the nonspecificity of clinical and biochemical criteria, strict histologic criteria for recurrence were established in the early 1990s.2 Since then, there have been many reports debating whether these findings indicate true recurrence.3–6 Donor age, recipient age, and warm ischemia time are possible risk factors for recurrence.7 The effect of immunosuppression on recurrence is debatable.3, 8–10

For PSC, recurrence was first suggested by Lerut et al.11 Thereafter, many scientific reports were published on disease recurrence in allografts. Harrison et al.12 reported that fibro-obliterative lesions were observed only in patients who underwent transplantation for PSC. Similarly, cholangiographic features suggestive of PSC were reported more often in transplants for PSC than in transplants for the other diseases.13 The standard criteria for diagnosis of PSC in the native liver were adopted to define disease recurrence.14, 15 Nonspecific bile duct injuries and strictures caused by allograft reperfusion injury, ischemia, rejection, and recurrent biliary sepsis should be excluded, because they might confound the diagnosis of recurrence.16–18 Type of immunosuppression has not been reported to influence the incidence of recurrence.18, 19

Recurrence of AIH in the allograft was first reported in 1984.20 Subsequently, many articles have been published on recurrence of AIH in allografts. However, there are no standard criteria to diagnose such recurrence. Many authors use the criteria of increased concentrations of transaminases or autoantibodies, hyper-γ-globulinemia, lobular or periportal hepatitis (or both) in the absence of rejection or viral infection, and steroid dependence.21–23 Human leukocyte antigen (HLA) phenotype, disease type and severity in native liver, level of immunosuppression, and duration of follow-up have also been reported as possible risk factors for recurrence.

While there have been many reports of recurrence of autoimmune liver diseases after liver transplantation, the exact incidence of recurrence and the factors contributing to it remain topics of discussion. Reported incidence of recurrence varies greatly among different studies. We conducted a systematic review of the English-language literature, followed by a pooled analysis, to determine the overall incidence of disease recurrence in allografts.


AIH, autoimmune hepatitis; CI, confidence interval ; HLA, human leukocyte antigen; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis.


We used MEDLINE (1966 to January 2006), EMBASE (1988 to January 2006), and Cochrane Database of Systematic Reviews to search for relevant articles. We used the following search strategy in the MEDLINE database: exp liver transplantation, exp recurrence/, exp postoperative complications, exp liver cirrhosis, biliary/or primary biliary cirrhosis. mp, exp cholangitis, sclerosing/or primary sclerosing cholangitis. mp, exp hepatitis, autoimmune/or autoimmune hepatitis. mp, exp autoimmune diseases/and exp liver diseases.

We also conducted a manual search of the reference lists in the review articles. Overall, 238 articles were identified. Two authors (M.G., V.B.) reviewed the abstracts of these articles to identify 90 potentially relevant articles (Fig. 1). Only original articles meeting inclusion criteria were used for the final analysis. In the case of studies using the same cohort, the most recent study was used.

Figure 1.

Flow chart shows final selection of published reports of studies on PBC, PSC, and AIH.

Data abstraction was done by one author (M.G.) who was not blinded to authors, affiliations, methods, or results. Selected studies were assessed for the total number of transplants, the number of persons with the disease in question, follow-up after transplantation, criteria for diagnosing recurrent disease in the allograft, number of recurrences, timing of recurrence after transplantation, nature of primary immunosuppression, and information on proposed risk factors for recurrence. Any conflicts in data abstraction were arbitrated by discussion with the senior author (V.B.). As most of the studies used for final analysis were retrospective observational studies, their quality could not be rated.

For PBC, original studies were included if they concerned adult patients who were 90 days past their liver transplantation, had histologic criteria diagnostic of recurrence, and had other causes of liver disease excluded. Histologic criteria included either the presence of florid duct lesions or 3 of 4 portal tract lesions (mononuclear inflammatory infiltrate, lymphoid aggregates, epithelioid granulomas, or bile duct damage).

For PSC, cohort studies were included if they concerned adults who were 90 days past their liver transplantation and had either radiologic or histologic criteria or both that were diagnostic of recurrence after other causes of similar radiologic and histologic features were excluded. Radiographic criteria were the presence of nonanastomotic biliary strictures, mural irregularities, and outpouching. Histologic features suggestive of recurrence were fibrosing cholangitis and fibro-obliterative duct lesions.

For AIH, original studies were included if they concerned patients who were 90 days past their liver transplantation and had histologic criteria diagnostic of recurrence with other causes of similar histologic findings excluded. The histologic criteria for recurrence were infiltration of portal tracts by plasma cells, piecemeal necrosis, and bridging necrosis.

Statistical Analysis

A pooled analysis produced a cumulative incidence of recurrence by combining the incidence reported in individual studies. Weighted incidence rates were also calculated because the number of subjects varied greatly among studies. Crude data on proposed risk factors were presented as percentages, and their 95% confidence interval (CI) was calculated. Crude data on proposed risk factors were combined for a calculation of the summary effects and 95% CIs. A funnel scatterplot was drawn to assess publication bias (i.e., selective reporting of large studies or studies featuring positive [statistically significant] results).



Thirty-five publications on PBC recurrence were identified (Table 1). Of them, 16 were reports of original studies (14 positive, 2 negative) that were used to determine the incidence of recurrence after liver transplantation.3–5, 8, 9, 24–34 Twelve other reports of studies that were identified were not included because of cohort duplication with later studies.1, 2, 7, 35–43 The prevalence of PBC in these studies was calculated to be 13%. The median age of patients at the time of transplantation was 52 years (range, 46.2-56 years). Most of the patients were women (90%). The median posttransplantation follow-up was 69 months (range, 36-114 months), and the median time to PBC recurrence was 46.5 months (range, 25-78 months). Of 1,241 patients with PBC, 204 (16%) had recurrence after liver transplantation. The weighted recurrence rate was calculated to be 18%. All but 2 publications reported recurrence of PBC in allografts.5, 24

Table 1. Primary Biliary Cirrhosis Studies
Original Articles
No.Author (reference)YearPBC, no.Recurrence, no.Age*MaleFemaleImmunosuppression
  • Abbreviations: NA, not available; MMF, mycophenolate mofetil.

  • *

    Age at liver transplant.

1Demetris et al.241988106046.21393All cyclosporine
2Gouw et al.5199419049019Cyclosporine or azathioprine
3Knoop et al.25199647950.5NANACyclosporine or tacrolimus
4Yoshida et al.261997161NANANAAll cyclosporine
5Slapak et al.271997338NANANAAll cyclosporine
6Sebagh et al.28199869651663All recurrences on tacrolimus; the rest on either tacrolimus or cyclosporine
7Hashimoto et al.29200162NA06All tacrolimus
8Renz and Ascher30200258856659All cyclosporine
9Yusoff et al.31200218254NANANA
10Salguero et al.322003196NA217Cyclosporine (17), tacrolimus (2)
11Levitsky et al.82003447NA638Cyclosporine (19), tacrolimus (22), azathioprine (10)
12Sylvestre et al.4200310017NANANANA
13Sanchez et al.9200316917/153NA8148cyclosporine ± azathioprine (71), cyclosporine ± MMF (49), tacrolimus (36)
14Kurdow et al.332003183/1752117Tacrolimus (1 recurrence), cyclosporine (2 recurrences)
15Neuberger et al.320044851145558427Cyclosporine (78 recurrences), tacrolimus (36 recurrences)
16Guy et al.342005564 of 48 (definite)53345Cyclosporine or tacrolimus
Duplicate Cohorts
17Neuberger et al.1198211350NANAAll azathioprine/MMF
18Esquivel et al.351988760NA769All cyclosporine
19Buist et al.361989520NANANANA
20Polson et al.371989239 of 10 biopsiesNA419All cyclosporine
21Samuel et al.38199051049546NA
22Hubscher et al.219938313NANANAAll cyclosporine
23Balan et al.39199360549555Cyclosporine (4 recurrences), tacrolimus (1 recurrence)
24Dmitrewski et al.401996271752.5423NA
25Van de Water et al.411996388NA038All cyclosporine
26Mazariegos et al.421997132NANANARestarted cyclosporine after initial weaning
27Liermann Garcia et al.72001400685440360NA
28Khettry et al.432003438NANANATacrolimus (13), cyclosporine (25)

Of 161 patients who received tacrolimus, 48 (30%) had recurrence of PBC after transplantation (95% CI, 23-38%). Of 430 patients who received cyclosporine, 98 (23%) had recurrent disease (95% CI, 19-27%). Whenever raw data were available, we calculated the relative risk of PBC recurrence with use of tacrolimus vs. use of cyclosporine for primary immunosuppression (Fig. 2).3, 8, 9, 43 While there was a trend toward more recurrence with tacrolimus immunosuppression compared to cyclosporine, it was not statistically significant (95% CI, −7.03- to 8.75%). We found no evidence of publication bias among the reports of the PBC studies (Fig. 3).

Figure 2.

Forrest plot shows reported relative risk of development of recurrence of PBC with tacrolimus vs. cyclosporine for immunosuppression in transplant recipients.

Figure 3.

Funnel plot illustrates lack of publication bias among reports of studies on PBC.


Thirty publications on PSC recurrence were identified (Table 2). To determine the rate of PSC recurrence after liver transplantation, 14 reports that emanated from original studies (all positive) were used.16, 17, 19, 30, 31, 44–52 Four other identified publications were excluded because their study cohorts were duplicated in later studies.12, 53–55 Prevalence of PSC was calculated to be 10%. The median age at the time of liver transplantation was 43.5 years (range, 36-53 years). There were more men (64%) than women (36%). Median follow-up after transplantation was 58 months (range, 53-77 months). Information on the timing of disease recurrence in allograft was missing in most of the studies. Of 940 patients who underwent transplantation for PSC, 161 (17%) had recurrence. The weighted recurrence rate was calculated as 11%.

Table 2. Primary Sclerosing Cholangitis Studies
Original Articles 
No.Author (reference)YearPSC, no.Recurrence, no.Age*MaleFemaleImmunosuppressionIBD Yes/No
  • Abbreviations: IBD, inflammatory bowel disease; NA, not available.

  • *

    Age at transplant.

1Marsh et al.451988551363520NANA
2Goss et al.44199712711478740Cyclosporine or tacrolimusNA
3Jeyarajah et al.17199811818NA7243All cyclosporine84/115
4Graziadei et al.46199912024458862NA94/109
5Kubota et al.471999533432726NANA
6Liden et al.482001615433922NANA
7Renz and Ascher302002497533118All cyclosporineNA
8Vera et al.49200215256NANANACyclosporine or tacrolimusNA
9Yusoff et al.31200212242NANANANA
10Khettry et al.162003516NANANANANA
11Kugelmas et al.1920037115445813Cyclosporine or tacrolimus55/71
12Brandsaeter et al.5120054994954Cyclosporine + azathioprine or cyclosporine + tacrolimusNA
13Khuroo et al.50200552NANANACyclosporine or tacrolimusNA
14Oldakowska-Jedynak et al.52200617236107Cyclosporine + azathioprine (7), tacrolimus (10)NA
Duplicate Cohorts
15Harrison et al.12199422344166Cyclosporine (20 recurrences), tacrolimus (2 recurrences)14/22
16Narumi et al.531995334NA1914NA22/33
17Saldeen et al.541999475432819NANA
18Liden et al.552000474432918NANA

We could not calculate the relative risk of recurrence with use of tacrolimus vs. cyclosporine as the primary immunosuppressant medication, because there was not enough information on use of immunosuppression in the individual studies. We also could not analyze various proposed risk factors (e.g., presence of native colon, history of colonic malignancy, or cytomegalovirus infection) because of inadequate information. Of 269 patients with inflammatory bowel disease, 48 (18%) had recurrence of disease (95% CI, 14-23%). Of 91 patients without inflammatory bowel disease, 16 (18%) had recurrence (95% CI, 11-27%). The relative risk of recurrence with inflammatory bowel disease was calculated whenever raw data were available.12, 17, 19, 46, 53 There was no statistically significant difference in the rate of recurrence between the 2 groups (95% CI, −4.49 to 6.03%). There was suggestion of publication bias among PSC studies (Fig. 4).

Figure 4.

Funnel plot illustrates publication bias among reports of studies on PSC.


Twenty-five publications on AIH recurrence were identified (Table 3). Thirteen original articles (12 positive and 1 negative) were analyzed.21–23, 31, 56–64 Three articles were excluded because of cohort duplication.30, 65, 66 Prevalence of AIH was calculated to be 3%. Median age at transplantation was 35.8 years (range, 19.5-44.5 years). Most patients were women (84%). Median posttransplantation follow-up was 43.8 months (range, 26.4-74.4 months). Disease recurrence was diagnosed in 94 (23%) of 414 transplant recipients after a median interval of 26.4 months (range, 14.4-55.2 months). The weighted recurrence rate was calculated to be 22%.

Table 3. Autoimmune Hepatitis Studies
Original Articles
No.Author (reference)YearAIH, no.Recurrence, no.Age*MalesFemalesANAAnti-SMAAnti-LKMGamma Globulin (g/L)Immunosuppression
  1. Abbreviations: ANA, antinuclear antibody; SMA, smooth-muscle antibody; LKM, liver-kidney microsomal antibody; NA, not available. *Age at transplant.

1Wright et al.5619924311NANANA68NAElevatedAll cyclosporine
2Prados et al.21199827934.642317117NATacrolimus (1), cyclosporine (26)
3Narumi et al.22199940537.91030NANANANATacrolimus or cyclosporine
4Milkiewicz et al.2319994713411136NANANA16.5NA
5Reich et al.572000246372302324NA3All recurrences on tacrolimus
6Ayata et al.58200012537.5210661ElevatedTacrolimus (5), cyclosporine (7)
7Gonzalez-Koch et al.592001417395NANANANANATacrolimus (4), cyclosporine (30)
8Molmenti et al.602002551144.51045NANANANATacrolimus (14), cyclosporine (43)
9Yusoff et al.31200212232NANANANANANANA
10Heffron et al.61200252932.11240NANANANANA
11Cattan et al.62200216019.5016NANANANAAll cyclosporine
12Duclos-Vallee et al.63200317730017914423.2All cyclosporine
13Vogel et al.6420042892942414166NATacrolimus (8), cyclosporine (19)
Duplicate Cohort
14Sanchez-Urdazpal et al.661992220NANANA1521NA2.6NA
15Ratziu et al.651999153231141010526.5All on cyclosporine
16Renz and Ascher302002371239928NANANANAAll on cyclosporine

Of 39 patients receiving tacrolimus, 12 (31%) had recurrence of disease (95 CI, 19-47%). Of 116 patients taking cyclosporine, 26 (22%) had recurrence (95% CI, 16-31%). We calculated the relative risk of AIH recurrence with tacrolimus vs. cyclosporine for immunosuppression whenever raw data were available (Fig. 5).58–60, 64 There was no statistically significant difference in recurrence between the 2 immunosuppressant medications (95% CI, −3.82 to 5.74%). There was suggestion of publication bias among AIH studies (Fig. 6).

Figure 5.

Forrest plot shows reported relative risk of development of recurrence of AIH with tacrolimus vs. cyclosporine for immunosuppression in transplant recipients.

Figure 6.

Funnel plot illustrates publication bias among reports of studies on AIH.


Liver transplantation is a well-accepted treatment modality for patients with end-stage liver disease secondary to autoimmune liver disease. As more liver transplant recipients survive longer, the recurrence of disease is becoming the primary cause of morbidity and mortality. Although recurrence of diseases such as viral hepatitis is fairly well accepted, true recurrence of autoimmune liver disease (e.g., PBC, PSC, or AIH) has been a subject of debate for a long time because of difficulties encountered in diagnosing recurrence.

After liver transplantation, patients with recurrent, early-stage autoimmune liver disease are asymptomatic (as they are in the pretransplantation course of the disease). Abnormal findings on liver tests may not be specific enough to identify disease recurrence, and autoimmune markers will frequently endure after liver transplantation; hence, the diagnosis of recurrent autoimmune liver disease can be challenging. Furthermore, posttransplantation pathologic processes such as rejection, graft-versus-host disease, and bile duct complications can complicate the diagnosis of these autoimmune liver diseases, which may be the reason for the controversy in the medical literature about their diagnosis. Different authors have reported different experiences with recurrence in the allograft; hence, the purpose of this systematic review is to provide some clarity on the issue. We observed that most studies reported recurrence of disease in the allograft. Pooled incidence rates showed that AIH recurred more frequently than PBC or PSC. Thus, we looked for publication bias in reporting recurrence. Publication bias is the selective reporting of large studies or those featuring positive or statistically significant results. If its presence is not detected, such bias can lead to misleading conclusions. The estimated effect may be larger, thereby resulting in skewed conclusions. There was suggestion of publication bias among PSC and AIH studies.

PBC Recurrence

In general, most histologic features of PBC after transplantation (e.g., bile duct paucity, ductular proliferation, and portal inflammation) may not be specific except for the presence of granulomatous bile duct destruction or florid duct lesion. To make the diagnosis of recurrent PBC, there should also be an absence of any acute or chronic rejection, graft-versus-host disease, or biliary obstruction. All but 2 PBC studies reported recurrence.5, 24

Levitsky et al.8 reported that recurrent PBC developed in the allografts of 7 of 46 patients. They did not report any difference in posttransplantation follow-up, incidence of acute rejection episodes, or rate of corticosteroid therapy between patients with or without recurrence. All patients with PBC recurrence were alive at the 3- and 5-year follow-up points. Sylvestre et al.4 compared 100 patients who underwent transplantation for PBC with 35 patients who received transplants during the same period for other conditions less likely to recur. Florid duct lesions were identified by protocol biopsies in 14 PBC patients, and 3 other PBC patients had destructive lymphocytic cholangitis within a dense portal infiltrate. Neither of these findings was noted in controls. Elevated alkaline phosphatase was noted in 8 (57%) of the 14 patients at diagnosis of the first recurrent disease lesion.

Neuberger et al.3 followed up 485 PBC patients a median of 79 months after transplantation. Protocol biopsies were done annually until 1997; however, the frequency of biopsies was not clearly stated for patients with liver transplants after 1997. Neuberger et al. found histologic evidence of disease recurrence in 114 patients. Guy et al.34 conducted 1 year or more of follow-up of 48 patients who had transplantation for PBC. Twenty-seven patients had elevated serum alkaline phosphatase. Biopsy specimens revealed definite recurrence in 4 (8%) of the 48 patients, probable recurrence in 11 patients (23%), and nondiagnostic abnormalities consistent with recurrence in 2 (4%).

Demetris et al.24 reported in 1988 that definitive diagnosis of recurrence was not possible after retrospectively reviewing histopathologic findings. Their study was done before the histologic criteria for recurrence were established in 1993.2 The criteria they used for recurrence were not clearly stated in their analysis. Biopsies were requested at the discretion of treating physicians on the basis of clinical or laboratory abnormalities, and protocol biopsies were not performed. Gouw et al.5 also did not report recurrence of PBC in their patients. They identified portal granuloma and mild bile duct damage by third-year liver biopsy in 2 out of 19 patients. These changes showed no evidence of progression in subsequent 3-year and 4-year follow-up biopsies, respectively.

Most published reports gave no information on the follow-up of transplant patients. The longest follow-up (median, 96 months) was by Sebagh et al.,28 who reported a recurrence in 6 out of 69 patients (9%). The shortest follow-up (median, 35.5 months) was reported by Hashimoto et al.29 Out of 6 patients, 5 were noted to have liver biopsies. Two patients with reported recurrence had follow-up for 36 and 50 months, respectively. In most reports, the onset of recurrence was reported within the first 5 years after transplantation. Sylvestre et al.4 identified florid duct lesions more than 7.2 years after transplantation in 3 of 14 patients with recurrence.

Role of Immunosuppression in PBC Recurrence

Immunosuppression may influence PBC recurrence in the allograft. There are conflicting reports on incidence of recurrence with either cyclosporine- or tacrolimus-based immunosuppression.7, 40 Mazariegos et al.42 reported PBC recurrence after rapid weaning from immunosuppressant medication. Levitsky et al.8 reported no difference in the rate of recurrence between patients maintained on cyclosporine vs. those taking tacrolimus. Sanchez et al.9 divided their cohort into 3 groups on the basis of immunosuppression: (1) cyclosporine plus azathioprine and prednisolone; (2) cyclosporine with or without mycophenolate mofetil and prednisolone; and (3) tacrolimus with and without mycophenolate mofetil and prednisolone. They reported no difference in recurrence rates among these groups. A more recent article has reported type of calcineurin-inhibitor used as a risk factor for recurrence of disease in the allograft as identified by multivariate analysis.3 The median time to recurrence was significantly shorter for the tacrolimus group than for the cyclosporine group (62 months vs. 123 months, P < 0.001). Additionally, Jacob et al.67 reported no difference in patient survival for cyclosporine- vs. tacrolimus-based immunosuppression. This study did report 1 recurrence of PBC in the cyclosporine group, but the diagnostic criteria were not clear.

Reports emanating from individual studies gave limited information on the use of immunosuppression. Nevertheless, we were able to use data from 4 studies and found an increased frequency of PBC recurrences in patients taking tacrolimus.3, 8, 9, 43 In the studies with no reported recurrence, primary immunosuppression was based on cyclosporine or azathioprine or both.5, 24 Clearly, the issue of immunosuppression needs further evaluation. We would recommend a multicenter, randomized study of PBC patients undergoing liver transplantation that would examine the effect of immunosuppression on these patients.

HLA Antigen Matching and PBC Recurrence

The presence of at least a single HLA antigen match between donors and recipients has been reported in patients who had a recurrence of PBC in their allograft.39 There was not enough information to study the extent of the effect of HLA antigen mismatches on PBC recurrence in individual studies. Sanchez et al.9 did not report any statistically significant difference in mismatches between the 2 groups. However, they did find significant allele similarities in donors in the recurrence group (P < 0.05). Hashimoto et al.29 reported their experience with living donors and did not observe any differences in HLA antigen mismatches between recurrence and nonrecurrence groups. Gouw et al.5 reported that most of their recipients (70%) shared variable numbers of HLA antigens with their donors, but they did not report any recurrence of disease in their patients. One might speculate that there is a lower rate of HLA antigen mismatch among patients in whom recurrent PBC develops, given the immune nature of the disease.

PSC Recurrence

Recurrence of PSC in patients after liver transplantation can be challenging because of the lack of a diagnostic standard criteria. In the native liver, the PSC diagnosis is based on the presence of bile duct strictures. However, in posttransplantation patients, this finding may be nonspecific and may be due to other reasons, such as allograft reperfusion injury, rejection, ischemia, biliary sepsis, ABO incompatibility, or technical complications. Hence, the diagnosis of PSC recurrence requires the presence of cholangiographic findings (as in the nontransplant setting), the presence of histologic findings of fibrous obliterative lesions, and the exclusion of other causes of biliary strictures. Histopathologic findings of recurrence may be indistinguishable from those of chronic rejection.17

Recurrence of PSC was reported in all studies, but most of the studies did not provide details on duration of follow-up and onset of recurrent disease in allograft. When Khettry et al.16 retrospectively reviewed posttransplantation liver biopsies of 51 PSC patients, they found histologic evidence of recurrence in 6 patients that included characteristic periductal fibrosis (3 patients), lymphoplasmacytic cholangitis involving the large duct (2 patients), and lymphoplasmacytic lobular hepatitis (1 patient). Of 6 patients with recurrent disease, 5 had cholangiograms that showed multiple nonanastomotic strictures involving the intrahepatic biliary tree. Kugelmas et al.19 reported PSC recurrence in 15 of 71 patients on the basis of cholangiographic findings (12 patients) or histologic features (3 patients).

So far, the largest series of patients with PSC recurrence has been reported from the University of Pittsburgh. Abu-Elmagd et al.68 presented their experience with 303 patients who had undergone transplantation for PSC over 16 years. They reported 52 recurrences in the allografts on the basis of typical cholangiographic and histologic features. We have not included this abstract in our analysis since this data is not yet published.

Cholangiographic features suggestive of recurrence have been published in 1996.13 More recently, Brandsaeter et al.51 compared magnetic resonance cholangiography findings for 49 patients who underwent transplantation for PSC with those for 45 randomly selected controls who received transplants during the same time for other causes. Magnetic resonance cholangiography showed the presence of bile duct strictures suggestive of recurrence in 9 cases and 1 control. Oldakowska-Jedynak et al.52 also reported PSC recurrence on the basis of magnetic resonance cholangiography findings in 2 patients in their cohort.

Marsh et al.45 reported that chemical and radiologic abnormalities compatible with recurrence developed in 1 of 55 patients 1 year after transplantation. However, the abnormalities did not progress during the following year, so diagnosis of recurrence was unclear.

Risk Factors for PSC Recurrence

Proposed risk factors for recurrent PSC include inflammatory bowel disease, cold ischemia time, number of cellular rejections cellular rejection episodes, previous biliary surgery, cytomegalovirus infection, and lymphocytotoxic cross-match. We could not analyze all the risk factors possibly associated with recurrence because of the limited information in individual articles. Khettry et al.16 found that HLA antigen class I matches were more frequent in patients with PSC recurrence (60%) compared with those in patients with recurrence of autoimmune liver disease, not otherwise specified (13%), and in patients without recurrence (30%). Brandsaeter et al.51 reported no difference in frequency of native HLA B8DR3 haplotype among PSC patients with recurrent disease and those without bile duct strictures. In their multiple regression analysis, steroid-resistant rejection was found to be a significant predictor of recurrence (P = 0.004; odds ratio, 26.3; 95% CI, 2.8-248.0). Vera et al.49 identified male sex and presence of intact colon before or after liver transplantation to be independent risk factors for disease recurrence in their multivariate analysis. Kugelmas et al.19 reported significantly (P = <0.01) shorter length of time for appearance of recurrence in patients with irritable bowel disease than in those without it. In our review, we did not find any significant difference in recurrence with the presence or absence of irritable bowel disease.

Role of Immunosuppression in PSC Recurrence

The limited data provided by individual studies precluded analysis of the association between type of primary immunosuppression and recurrent PSC. Primary immunosuppression with either cyclosporine or tacrolimus did not seem to influence recurrence in 2 studies.12, 19 Kugelmas et al.19 reported a higher risk of PSC recurrence with OKT3 therapy.

AIH Recurrence

The diagnosis of recurrent AIH is particularly challenging after transplantation, because the presence of periportal hepatitis, with or without lobular hepatitis, can be a nonspecific finding. Markers that may be helpful in the pretransplantation diagnosis of AIH, such as the presence of elevated transaminases associated with hyper-γ-globulinemia, the presence of autoantibodies, and the absence of other causes of chronic liver disease may persist after transplantation.

Recurrence of AIH after liver transplantation has been reported by most transplant centers. Currently, there are no defined criteria for diagnosis of recurrence. In the reports we reviewed, all but one included studies that reported recurrence.62 Molmenti et al.60 reported 11 recurrences after prospectively collecting data on 55 AIH patients who underwent transplantation. One patient experienced 2 recurrences. Duclos-Vallee et al.63 reviewed records of 17 AIH patients with more than 10 years of posttransplantation follow-up. Histologic recurrence was reported in 7 patients (41%). In these 7 patients, recurrence was diagnosed on the basis of protocol biopsies in 4 patients and by histologic findings of liver test abnormalities in 3 patients.

Cattan et al.62 did not report recurrence of AIH after 6 years of median follow-up of their cohort of 16 patients. Three patients (19%) underwent emergent transplantation for acute liver failure. These 3 and 1 other patient (25%) died during the first 3 months after transplantation. Sanchez-Urdazpal et al.66 also did not report recurrence. A subsequent study on the same cohort, however, did report recurrence of AIH in allograft.59

Risk Factors for AIH Recurrence

High-grade inflammation in the native liver has been reported as a strong predictor of AIH.58 Some authors have reported an increased frequency of the HLA-DR3 haplotype in patients with AIH recurrence,22, 59 whereas others have not observed any such differences.60, 64 Wright et al.56 reported that recurrence of AIH was more likely to occur in HLA-DR3-positive recipients of HLA-DR3-negative grafts. However, this particular HLA antigen mismatch was not observed in a subsequent study.59 HLA antigen mismatch was not found to differ significantly in patients with AIH recurrence.23, 58, 63, 65 Total numbers, incidence before recurrence, and treatment of acute cellular rejection were not reported to differ significantly between patients with or without AIH recurrence.60

Role of Immunosuppression in AIH Recurrence

We found no difference in AIH recurrence with use of cyclosporine vs. use of tacrolimus for immunosuppression. Molmenti et al.60 reported no difference in recurrence with respect to initial immunosuppression. They also did not observe any difference in pretransplantation corticosteroid administration between patients with or without recurrence. Heffron et al.61 reported no difference in recurrence between patients receiving steroids for different durations. Of the 9 patients with recurrence (17%), only 3 had been weaned completely from steroids. Since there was limited information on steroid duration and withdrawal in individual studies, we did not explore this association.

In summary, our systematic review of the literature supports the recurrence of autoimmune liver diseases after liver transplantation. As patients are followed longer, recurrence of disease may become the primary cause of morbidity in these patients in the long term. For PBC, the presence of classic histologic features, such as granulomatous duct destruction and exclusion of other biliary processes, is helpful to diagnose recurrence. In contrast, recurrent PSC has been more controversial. Nevertheless, it is increasingly accepted, especially in the presence of well-defined cholangiographic and histologic criteria and in the absence of other processes such as preservation injury or blood ABO incompatibility or in the absence of hepatic arterial problems. Lastly, recurrence of AIH is based on a combination of clinical and biochemical findings and on the absence of any viral infection, drug effects, or any other causes of chronic liver disease.

Most studies were uncontrolled retrospective studies with different study designs. Hence a limitation of our analyses is that the methodological quality of studies could not be assessed. A meta-analysis of controlled studies should have been the ideal way to determine the incidence of recurrence in the allograft. Meta-analysis is a form of critical review of literature on a stated subject with the emphasis on producing quantitative conclusions.69 For example, a meta-analysis of randomized controlled trials would be the ideal way to determine the influence of primary immunosuppression on recurrence. In the absence of such investigation, this systematic review provides a critical assessment of the evidence currently available to us. Furthermore, the incidence of graft loss due to recurrence of these diseases remains unknown from these studies. Different authors have previously reviewed the recurrence of autoimmune liver diseases after liver transplantation (Table 4). However, as these were narrative reviews, a comparison with these previously published review articles could not be done. Furthermore, not all reviews have looked at all 3 diseases together.

Table 4. List of Review Articles
No.Author (reference)Year
2Batts et al.711998
3Balan et al.721999

Differences in study design, immunosuppressive regimens, and follow-up have made the systematic review challenging. Hence, we recommend that the following information be included in future manuscripts reporting on recurrence of autoimmune liver diseases in the allograft: (1) clinical and histological criteria used to diagnose recurrence; (2) HLA matching information; (3) complete demographic information at transplant; (4) for PSC studies, information regarding the presence and duration of inflammatory bowel disease, history of colorectal cancer and colectomy; (5) criteria used to exclude diseases that can mimic by histology or radiology recurrent disease; and (6) primary immunosuppression used and the use of steroids and their duration.

We would like to propose the following guidelines for the long-term posttransplant follow-up of these patients. (1) Biochemical markers such as antimitochondrial antibody, antinuclear antibody, and quantitative immunoglobulin levels are obtained in PBC and AIH patients on an annual basis. (2) Histological criteria listed in the methods section of this manuscript are used for diagnosing recurrence. (3) Radiological criteria listed in the methods section are used for diagnosing recurrent PSC. (4) Liver biopsies are obtained at posttransplant years 3 and 5, and thereafter at 3-5 year intervals. In addition, multicenter studies are needed to address the role of immunosuppression on recurrence of autoimmune liver diseases in the allograft when patients are randomized to receive either cyclosporine or tacrolimus along with a specified steroid protocol. These studies should be designed to obtain at least 5-10 years of follow-up, as doing so will also provide information on the natural history of autoimmune liver diseases after liver transplantation.


The authors thank Kay E. Wellik for library support, Bonnie L. Schimek for support with figures and illustration, and Mary Ann Clifft for editorial assistance in preparation of this manuscript.